You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The server running in the docker environment does not respond correctly to the shutdown request, but it is killed at the end of the timeout. A longer timeout produces no advantages.
This behavior also causes side problems, including permanent errors in connectors when the connected service (e.g. mqtt broker) is running on the same server, and the server is being stopped. See also #771#772.
The issue can be verified by stopping the Lorawan container and measuring the shutdown time, which will be approximately 10 seconds (corresponding to the timeout for stopping the containers).
I know that this project is no longer being maintained for now, and I hope that it will be maintained again in the future. Best regards to @gotthardp for the excellent project.
The text was updated successfully, but these errors were encountered:
lcse66
added a commit
to lcse66/lorawan-server
that referenced
this issue
Nov 25, 2023
The lorawan-server script doesn't forward TERM signal to underlying process.
Normally the underlying process must be started from the starting script with "exec", but only if it is a binary executable (which "erl" is not; from tests carried out it works, but two "zombie" processes remain active). In this modification, a trap is activated that forwards the term signal.
In solution to gotthardp#843 issue
Hi Hugo (@hugosantosbessa)
Sorry but I didn't understand the problem well. The mqtt connector works fine for me. The only problem is (was) that during the host shutdown the Lorawan server closes late, after the mqtt broker has stopped, and this causes a permanent error on the connector which requires manual reset (deleting the "badarg" flag by the connector configuration). With the change #844 that I proposed, the problem never recurred.
Uh oh!
There was an error while loading. Please reload this page.
Hi everyone.
The server running in the docker environment does not respond correctly to the shutdown request, but it is killed at the end of the timeout. A longer timeout produces no advantages.
This behavior also causes side problems, including permanent errors in connectors when the connected service (e.g. mqtt broker) is running on the same server, and the server is being stopped. See also #771 #772.
The issue can be verified by stopping the Lorawan container and measuring the shutdown time, which will be approximately 10 seconds (corresponding to the timeout for stopping the containers).
I know that this project is no longer being maintained for now, and I hope that it will be maintained again in the future. Best regards to @gotthardp for the excellent project.
The text was updated successfully, but these errors were encountered: